Reader Comments

Post a new comment on this article

Moderator analysis request: Is there an effect of control-group quality?

Posted by Michael_J_Kane on 22 Dec 2014 at 14:18 GMT

Given that 50% of the studies included in the meta-analysis used an active control group, I hope you will consider testing whether this indicator of control-group quality acted as a moderator of training benefits. It would be worrisome if the strongest evidence for effectiveness came from studies using passive control groups. (Of course, even an active control group may not be effective as a placebo control, but active controls are clearly preferable over passive ones.)

Thank you for considering my question and request for further analysis.

Michael J. Kane
mjkane (at) uncg (dot) edu
University of North Carolina at Greensboro

No competing interests declared.

RE: Moderator analysis request: Is there an effect of control-group quality?

michaelvalenzuela replied to Michael_J_Kane on 25 Dec 2014 at 00:58 GMT

Hi Michael and thanks for your comment.

If you look at Figure 11 you will see that we tested for the moderating effect of control group - active vs passive. Effect size estimates from both subgroups were very similar and there was no significant interaction effect. Therefore, there is no evidence that type of control group has a systematic effect one way or another on cognitive outcomes in older individuals.

For your information, on a post hoc basis we have estimated the aggregate effects size from RCTs of computerised cognitive training in older non-demented individuals that used only supervised group-based training AND active controls. The results are in-line with those reported in our paper (if anything tending to be stronger) and highly consistent: k=11 RCTs, g=0.34 (95%CI 0.24 to 0.45), I^2=0%.


Michael Valenzuela

No competing interests declared.